Something different: Outliers, Artifacts, and Anecdotes

Some data does not fit the pattern we identify or the narrative that guides us. How people approach a datum that does not stand in line can be ... fuzzy. Grabbing a couple of  commonly shared thoughts on the matter: 

• trimming
• winsorization
• "the plural of anecdote is not data"
• "the exception proves the rule"

"You don't count because you're different," is not usually phrased that bluntly, but I have read or heard:

• "oh, you're (or they're) an outlier"
• "sampling 100 [soldiers, or college students] at [my community]"
• "I've never met someone like what you're saying"

On the other side of inclusion, we do see a couple of phrases show up:

• "Huh. That's funny." ("weird" may be the more up-to-date word)
• "Eureka!"
• "Hey, check this out..."

"The plural of anecdote is data" - Ray Wolfinger, and yes, his use of that at UC-Berkley to encourage people to be thoughtful about outliers got mangled into the now common "the plural of anecdote is not data."

Related links:

• https://ritholtz.com/2019/02/the-plural-of-anecdote-is-data/
• https://www.strataquant.com/post/how-anecdotal-evidence-can-make-or-break-your-insights
• https://blogs.iq.harvard.edu/the_singular_of